COVID-19 Health Disparities and Their Causes

COVID-19 did not hit everyone equally. Throughout the pandemic, Black, Hispanic/Latino, American Indian or Alaska Native, and Asian or Pacific Islander individuals in the United States were hospitalized, admitted to intensive care, and died at significantly higher rates than white individuals, and these disparities persisted across all age groups for the entire first year of the pandemic.¹ The causes were not primarily genetic or biological. They were structural: where people lived, where they worked, what insurance they carried, what chronic conditions they had already developed under unequal systems, and how the healthcare system treated them once they got sick. Understanding these overlapping causes matters because they did not disappear when the acute pandemic receded.

Where People Lived

Housing was one of the most direct pathways to unequal infection rates. Crowded living conditions make respiratory virus transmission almost inevitable, and certain communities were far more likely to live in tight quarters. A prospective cohort study in England and Wales found that overcrowded households had roughly two to four times the odds of confirmed SARS-CoV-2 infection compared with households that had more rooms than occupants, even after adjusting for other demographic and socioeconomic differences.² In the United States, the connection between density and transmission was even starker. Among households with more than three people sharing fewer than six rooms, the secondary attack rate reached about 91%, and the overall household transmission rate in minority households was 70% compared with 52% in white households.³

Housing quality mattered independently of crowding. A county-level analysis found that for every five-percentage-point increase in the share of households with poor housing conditions, COVID-19 incidence rose by about 50% and mortality by about 42%.⁴ Poor housing in this context includes factors like lack of complete plumbing, overcrowding, and high cost burden that forces multiple families to share space. These conditions are not randomly distributed; they cluster in communities shaped by decades of discriminatory lending, zoning, and disinvestment.

Where People Worked

The ability to work from home was perhaps the single sharpest line dividing pandemic risk, and that ability broke along racial and ethnic lines. Black workers were overrepresented in occupations with the highest potential exposure to disease and the least ability to maintain physical distancing. Hispanic workers were overrepresented in jobs with the lowest likelihood of remote work.⁵ These were not niche findings. A multi-city analysis found that about 70% of foreign-born Hispanic workers were classified as essential, compared with roughly 63% of non-Hispanic white workers, and the gap in high-exposure-risk work varied widely by city.⁶

The practical meaning of these numbers is straightforward: meatpacking workers, bus drivers, grocery clerks, and home health aides could not isolate themselves from the virus the way an office worker logging in from a spare bedroom could. This was a pandemic of proximity, and proximity to other people was determined more by your paycheck and your job than by your personal choices.

Chronic Conditions That Made Infections Worse

Once infected, some people were far more likely to end up in the hospital or on a ventilator, and the conditions that drove severe outcomes were themselves unevenly distributed. Poorly controlled high blood pressure raised the odds of hospitalization by about 30%, while poorly controlled diabetes raised hospitalization odds by roughly 60%.⁷ Black, Hispanic, and Indigenous communities in both the United States and the United Kingdom carry higher burdens of these conditions, a reality that predates the pandemic by generations.⁸ The diabetes and hypertension did not appear out of nowhere. They reflect lifetimes of unequal access to healthy food, safe exercise spaces, preventive care, and consistent health insurance.

This is where the biology-versus-society debate gets misleading. A person’s blood pressure is a biological measurement, but the reason it is high is usually a story about poverty, stress, food environments, and healthcare access. COVID-19 turned those upstream inequities into downstream death counts.

Insurance, Cost, and Getting Through the Door

Having health insurance does not guarantee good care, but not having it almost guarantees worse care. During the pandemic, about two-thirds of recently uninsured adults cited unaffordable coverage as the reason, and Hispanic adults who named cost as a barrier actually increased by about two percentage points after the pandemic began.⁹ In four Southern states, roughly 31% of low-income adults reported delaying care because of cost in 2020.¹⁰

There was some good news buried in the numbers: states that had expanded Medicaid saw uninsurance rates actually decline during the pandemic, with the reductions concentrated among Black, multiracial, and low-income residents.¹¹ This contrast between expansion and non-expansion states is itself evidence that policy choices, not just individual circumstances, determined who could get care.

Bias Inside the Healthcare System

Even after patients arrived at a hospital, the care they received was not always equal. Research consistently shows that healthcare providers carry both explicit and implicit biases against racial and ethnic minorities, and these biases affect clinical decision-making and patient communication.¹² The pandemic likely made this worse, not better. When hospitals were overwhelmed, providers were making faster decisions with fewer resources, exactly the conditions under which implicit bias tends to sharpen.¹³ Older, Black, uninsured, rural, and less-educated patients were disproportionately affected in both illness and death, and limited data on unconscious bias in healthcare suggests it consistently alters clinical decisions in ways that disadvantage these groups.¹⁴

Unequal Access to Vaccines and Treatments

When vaccines arrived, they did not reach everyone at the same pace. Across the U.S., about 16% of adults in lower socioeconomic status counties were vaccinated early on, compared with about 20% in higher-status counties. Among adults 65 and older, the gap was wider: roughly 38% versus 48%.¹⁵ Hispanic respondents had the largest gap compared with white respondents, about eight percentage points lower coverage, and socioeconomic factors such as insurance, income, education, and employment explained the majority of that difference.¹⁶

Treatments showed similar patterns. When monoclonal antibody therapy was available, only about 45% of Hispanic/Latino and 31% of Black patients received it, compared with 64% of white patients. After adjusting for age, insurance, language, illness severity, and neighborhood vulnerability, the racial gap itself was no longer statistically significant, but being uninsured or speaking a primary language other than English remained strong independent barriers.¹⁷ In England, a similar story emerged: Black patients were treated with antivirals and monoclonal antibodies at roughly half the rate of white patients, and patients in the most deprived areas received treatment at about two-thirds the rate of those in the least deprived areas.¹⁸ The takeaway is the same across the Atlantic: race and poverty predicted who got life-saving treatments, and the mediating factors were access, language, insurance, and geography, not biology.

Medical Mistrust and Its Roots

Some of the disparity in vaccine uptake was driven by hesitancy, but calling it “hesitancy” alone misses the context. For Black communities in particular, distrust of medical institutions is rooted in documented historical abuses and ongoing structural racism, and the pandemic did not create that distrust so much as expose it.¹⁹ Across Black, Asian, and Latino populations, both individual-level and group-level medical mistrust were significantly associated with lower intent to vaccinate.²⁰

Experiences of everyday discrimination, healthcare discrimination, and structural discrimination were all linked to higher general medical mistrust and to greater endorsement of COVID-19 conspiracy beliefs.²¹ In other words, the distrust was not irrational. People who had been treated badly by the healthcare system in the past were less inclined to trust that same system when it said “get vaccinated.” Framing this as a knowledge deficit or a cultural problem misses the point: it is a trust problem, and the institutions broke the trust.

Prisons and Congregate Settings

Perhaps nowhere were disparities more extreme than in prisons. By April 2021, the cumulative COVID-19 case rate in U.S. prisons was more than three times the rate in the general population, and the standardized mortality rate was about 2.5 times higher.²² During peak weeks, the prison-to-general-population infection ratio hit five to one. In Florida prisons, all-cause mortality in 2020 was 42% higher than the year before, and life expectancy at age 20 dropped by more than four years in a single year.²³

This connects directly to racial health disparities because the U.S. prison population is disproportionately Black, Hispanic, and Indigenous. Incarcerated people had little control over distancing, ventilation, or access to healthcare, and a systematic review confirmed that COVID-19 infection risk in prisons was consistently higher than in the general population worldwide.²⁴

Language Barriers and Information Access

Public health messaging during the pandemic was overwhelmingly produced in English and often used technical language that was difficult even for fluent English speakers. For people with limited health literacy or who spoke other languages, critical information about testing, isolation, and vaccination was often inaccessible. Among parents who were deaf or hard-of-hearing, those with limited health literacy had more than double the odds of reporting barriers to accessing COVID-19 information, and Hispanic respondents also had significantly higher odds of these barriers.²⁵

Qualitative research in Australia found that culturally and linguistically diverse communities faced persistent challenges in receiving timely, clear, and culturally appropriate public health messages from government and community organizations.²⁶ When you cannot understand the rules or where to get tested, your risk goes up, and these communication failures layered on top of every other structural disadvantage.

Neighborhood Vulnerability

The social vulnerability of a neighborhood, which bundles together poverty, crowded housing, minority status, language barriers, and limited transportation, was itself predictive of COVID-19 outcomes. In both Alabama and Louisiana, the most socially vulnerable communities had significantly higher case rates, with rate ratios of about 1.6 and 1.4 respectively compared with the least vulnerable areas.²⁷ Interestingly, the relationship between vulnerability and testing access differed between the two states: Louisiana saw more testing in vulnerable areas, while Alabama did not. Where testing was limited, the true case count was almost certainly undercounted, meaning the disparity was likely even larger than the data showed.

Pregnancy and Children

Disparities extended to pregnant women and children. Black pregnant patients were about twice as likely to be infected with COVID-19 compared with white patients, and this elevated risk persisted even after accounting for obesity and diabetes.²⁸ In Florida, higher county-level COVID-19 infection rates during the first trimester were associated with increased preterm birth and low birth weight for all women, but the effect was significantly more pronounced for Black women. Each one-percentage-point increase in COVID-19 cases was associated with a 1.2-percentage-point increase in preterm birth probability for Black women beyond the increase observed in white women.²⁹

Among children, multisystem inflammatory syndrome (MIS-C), a rare but severe post-COVID condition, showed striking racial patterns. CDC data reported that 31% of MIS-C cases occurred in Black non-Hispanic children and 26% in Latino children, far exceeding their share of the overall child population.³⁰ Both race/ethnicity and socioeconomic status independently raised the odds of a MIS-C diagnosis, with Black and Hispanic children and those in the lowest-income and highest-vulnerability neighborhoods at significantly elevated risk.³¹

Mental Health

The pandemic’s mental health toll was not shared equally either. In a large U.S. sample, Black, Hispanic, Asian, and multiracial respondents all had higher odds of screening positive for depression compared with white respondents, even after accounting for prior mental health history and COVID-19 infection status. The pattern was mirrored in a large UK sample.³² The mental health burden was compounded by events outside the pandemic itself: Black adults experienced especially high distress around the murder of George Floyd, and Asian adults around the Atlanta spa shootings. Meanwhile, white respondents were consistently more likely to receive professional mental health care both before and during the pandemic, meaning that the groups with the greatest need had the least access to support.³³

Long COVID and the Follow-Up Gap

The disparities did not end when the acute infection resolved. Women, people in deprived areas, those with learning disabilities, homeless individuals, and some minority ethnic populations have high rates of long COVID but low access to follow-up support, creating a clear inequity in ongoing care.³⁴ In the UK, people in the most deprived areas had higher risk of long COVID compared with those in the least deprived areas, and socioeconomic inequalities interacted with each other to worsen outcomes.³⁵

There is an additional, more insidious problem with long COVID data. Among patients hospitalized for COVID-related respiratory distress, Hispanic/Latino patients were more likely to be uninsured and discharged home rather than to rehabilitation. Since long COVID diagnosis depends on follow-up care, lack of outpatient access means many people are never diagnosed, and they end up underrepresented in prevalence estimates.³⁶ The disparity is not just in who gets long COVID but in who gets counted as having it.

Indigenous Communities

American Indian and Alaska Native populations were among the hardest hit. The incidence and severity of COVID-19 were disproportionately high in Native American communities, driven by a combination of chronic underfunding of the Indian Health Service, geographic isolation, limited infrastructure, and high rates of underlying conditions.³⁷ Even when Indigenous patients overcame barriers to reach care, they received worse treatment than white patients on an estimated 40% of quality measures. Systemic underfunding, discrimination, and an inability to see preferred clinicians were widely cited as barriers that the pandemic further exposed.³⁸

Community Health Workers as a Counterweight

One of the more effective responses to these layered disadvantages was the deployment of community health workers, people embedded in the communities they serve who could bridge language, trust, and access gaps. The CDC’s Community Resilience program hired about 950 community health workers and integrated them into over 1,000 organizations, resulting in more than 250,000 referrals to social services and 150,000 referrals for specific health conditions.³⁹ In rural western Kenya, a comparison between counties found that training and deploying community health workers was associated with a 46% reduction in COVID-19 infections and a 71% reduction in COVID-19 deaths.⁴⁰

These programs worked because they addressed multiple upstream problems at once: they provided information in the right language, built trust that institutional channels had eroded, connected people to testing and vaccination, and linked them to social services that kept them housed and fed. The evidence suggests that the most effective pandemic responses were not purely medical. They were structural, meeting people where they were and dismantling barriers one at a time.